UTM Parameters

The five tags and what each one answers

utm_source answers “where did the click originate?” (google, meta, newsletter). utm_medium answers “through what mechanism?” (cpc, email, organic-social). utm_campaign names the initiative (spring-sale-2026). utm_term carries the paid keyword, and utm_content separates variants of the same placement — two banners in one email, two creatives in one ad set. Source and medium are the load-bearing pair: they feed channel grouping, and a visit tagged with them lands cleanly in last-click attribution instead of the referrer-guessing fallback.

Why naming discipline is the whole game

UTM values are not validated by anyone. “Facebook”, “facebook” and “fb” become three different sources; “email” and “Email” split one channel’s revenue across two rows. The fix is organisational, not technical: one shared naming convention (lowercase, hyphens, an agreed vocabulary for source and medium), one link-building spreadsheet or tool, and the discipline to reject links that break the scheme. An untagged campaign link is worse than an inconsistent one — the visit falls back to referrer data, which browsers increasingly strip, and typically ends up inflating direct traffic where no budget decision can reach it.

Why UTMs survive the cookieless era

Most attribution mechanisms depend on state stored in the browser — cookies, local storage, device identifiers — and all of that is consent-gated in the EU and actively restricted by Safari and Firefox. UTM parameters carry the attribution in the URL of the visit itself, so nothing needs to be stored or read back later. That is why they are the backbone of cookieless analytics: SealMetrics reads the UTM set on each pageview and attributes conversions last-click at channel level, on 100% of visits, without a consent banner — while a cookie-based tool applies the same tags to only the minority of traffic it is allowed to observe.

What UTM parameters do not tell you

UTMs describe the click, not the person. They cannot connect this visit to an earlier one, credit an ad that was seen but not clicked, or follow a journey across devices — and within a session, they only describe how it started. They are also self-reported: a UTM says what you wrote on the link, not what the traffic actually was, so a mislabelled link lies with complete confidence. Treat them as the input to channel-level revenue attribution, not as a record of user behaviour.